On 5/24/08, Luís Oliveira luismbo@gmail.com wrote:
- A parenscript-like language for GLSL. See
http://advogato.org/person/ingvar/diary/166.html for an initial attempt at doing that.
Might be nicer to have something smarter than parenscript, so it could do type inference, and maybe some simple optimizations.
- Adapt the OpenGL documentation for cl-opengl. (Something like what
the Python bindings have, I suppose.)
Something that could hook into slimes documentation hotkeys (C-c C-d h, etc.) would be particularly nice :)
- Translate the rest of the redbook examples to cl-opengl/cl-glut.
(Not exactly a big priority, but it'd be nice anyhow.)
- Finish the bindings. While everything is available from the %gl
package (since that is autogenerated from the spec) not everything has been lispified yet.
I'd probably lean towards writing some examples of more modern OpenGL usage, particularly for someone who uses GL a lot already. Would also be a good way to prioritize which of the missing bindings would be useful to wrap further.
- Not exactly specific to cl-opengl, but something better (in
particular, more Lisp-friendly) than GLUT would be useful. I learned about LGLFW recently (http://bollhavet.ath.cx:1080/~mikezor/lglfw/) but I have not tried it. Porting cl-glut's CLOS-based API to that might be interesting.
see also http://www.cliki.net/cl-glfw , which claims to work with cl-opengl.
Examples of using cl-opengl with other alternatives to glut would also be good, lispbuilder-sdl, CLX, direct glx/wgl/agl/etc...
Performance testing/tuning would also be useful, see how close to C we can get, and figure out what sorts of things need work (like texture uploading doesn't look like it will be good until CFFI gets an API for sending lisp arrays to FFI without a copy).
-b-